The present disclosure relates to an inter-object relation recognition apparatus, a learned model, a recognition method, and a program for recognizing a relation between objects included in an image.
It is known that there is an inter-object relation recognition apparatus that learns a plurality of relations between objects included in an image, and recognizes and outputs the relation between objects, based on the learning result (See, for example, Yuval Atzmon, Jonathan Berant, Vahid Kezami, Amir Globerson, Gal Chechik, Learning to generalize to new compositions in image understanding, arXiv preprint arXiv; 1608.07639, 2016).